- Aug 24, 2021
- 208
- 613
LOL mate, quite the opposite.lmao speaking like someone who knows 0 about the subject.
You do realize there are different loras, checkpoints, styletransfers, embedding, hypernetoworks, loras, controlnets, VAEs, DoRAs, diffusion models, ipadapters, upscalers and first of all the possibility for endless prompting, right?
On top of that you have like 5 good base models in which you can work on and that have been exponentially getting better.
Dude sees an image probably done with a simple prompt on top of rundiffusion which isnt even a local stable diffusion where you have endless possibilities and is like "yeah AI Art is bad, despite these images being better than 99% of the renders in this thread and infinitely better than the monstruosity I have on my avatar that I call art".
Seriously?
Not to brag, but this is all my personal work, as I painted it myself, by hand, no AI whatsoever:
So yes, I do know what I am talking about and defiantly know way more than you.
Edit:
Here, 5 minutes of me correcting things:. Fixed her leg, dress, arse, lamp reflection on the water, added highlights, sky colour, fixed plastic skin, + her face but only slightly. Added subtle reds, yellow and greens tones to her skin so it looks natural. Fixed shadows and added where they were missing.
Still wrong - here wrist bone being on top of here forearm rather than side, elbow bone going sideway, left side of the water feature is higher than other
Last edited: